ABSTRACT
There is a high demand for the development of reliable chronic toxicity tests using tropical marine species for subsequent use in tropical risk assessment. However, many chronic test endpoints can be laborious and time-consuming to assess, particularly if the endpoints require measurements of individuals (e.g. growth, size) or advanced taxonomic expertise (e.g. differentiating between larval development stages). In this study, we used scanning and image recognition (SIR) technology to develop and validate a chronic toxicity test with larvae of the tropical euryhaline copepod, Acartia sinjiensis. Optimisation steps are described, and included egg age, and effect of algal food type and salinity on toxicity. Comparisons were made between traditional endpoints measured using microscopy and those measured using SIR. Traditional endpoints of larval development ratio (LDR) and survival achieved using microscope examination and SIR were almost identical (R2â¯=â¯0.96-0.97). Additional endpoints made possible by SIR included larval development index (LDI; based on the number of animals at different stages of development), and a range of size measurements (e.g. surface area, perimeter and length) for individual animals and for total populations (i.e. a proxy for biomass). The SIR-derived endpoints were based on measurements that had concentration-dependant responses to tested toxicants (copper, nickel, ammonia), and were a sub-set of the full range of metrics provided by the software. Toxicity values based on SIR-measurements were similar to or more sensitive than the traditional LDR endpoint. SIR technology provides a major opportunity to improve and modernise larval development tests for a range for species, but comes at a cost of increased data size and complexity. Therefore, as a research tool, SIR has significant advantages over traditional microscope methods, but for routine toxicity testing, SIR incorporation into invertebrate toxicity testing will benefit from further improvements to the associated software and data management systems.
